Table Design

Stat 365: Statistical Communication

Wednesday, April 24th1

Today we will…

  • The role of tables
  • Table design
  • Practice with tables

The role of tables

Use a table when…1

  • The display will be used to look up individual values
  • The display will be used to compare individual values rather than a series or trend
  • Precise values are required
  • Values involve more than one unit of measure
  • Values must be presented at various levels of aggregation

Table Anatomy1

Table Tips

  • Tables should be self-sufficient and stand-alone. The reader should be able to understand the table without reading your discussion. Use titles, row headings, and column headings to convey:
    • Purpose
    • Context
    • Variables (units of measurement/categories)
    • Data sources (in footnotes)
    • Definitions of technical terms
  • Let the data speak for themselves – Simplicity!
  • Organize material in a meaningful way
  • Tables should never flow across a page break

Table Example: Childcare Costs

Table X. Descriptive Statistics on ca_childcare
study_year facility_type avg_weekly_cost sd_weekly_cost
2008 mc 245.0569 36.12943
2018 mc 277.3488 70.83294
2008 mfcc 155.0660 27.43181
2018 mfcc 196.1191 41.84334
  • Short, cryptic acronyms
  • References to mc, mfcc, not available to your reader
  • Excessive digits
  • If we want to compare center vs family care we have to jump rows.
Table X. Means and Standard Deviations of the Median Weekly Cost for Infant Childcare across the 58 counties in California
Facility Year Mean ($) Std. Dev.
Center-based 2008 245.1 36.1
Center-based 2018 277.3 70.8
Family-based 2008 155.1 27.4
Family-based 2018 196.1 41.8
Note:
Data source: National Database of Childcare Prices (NDCP)
  • Labels clearly identify the concepts in each row
  • Ordering of rows makes comparison of center vs family care easy
  • The variable can be defined in the data and methods section or a note to the table
  • For a non-scientific article, replace technical labels with everyday synonyms

Table Design

Align numbers to the right

Select your font to be legible

Variations in tables

Unidirectional categorical items are laid out in one direction only (think long data)

Bidirectional categorical items are laid out in both directions (think wide data)

Highlight what’s important!


Boldfaced or Italic text

Size

Color
intensity
Position

Common Errors1

  • Tables that are too large or too simple
  • Inclusion of nonessential data
  • Failure to use shading and bordering in tables
  • Redundancy with text
  • Excessive precision
  • Not self-explanatory
  • Design elements interfere with the clarity
  • Inadequate definition of symbols or abbreviations

Packages for tables in R

  • library(gt)
  • library(DT)
  • library(kable)
  • library(kableExtra)

Practice in table design

Can you sketch out a table of summary statistics you might include in your final group report for your data?

To do

One-number Story

  • Work on improving your story
  • Canvas Submission due Thursday, 4/25

Listen to Storytelling with Data Podcasts (Data Viz)

  • For class next week